面向工业数据调和、参数估计与过程模拟的精馏系统一体化平台设计OA北大核心CSTPCD

Process data is the basis for simulation and optimization of chemical process systems.However,some important variables may not set measuring points for many production processes,which is not able to obtain,and the data that can be collected usually has poor reliability.Data reconciliation and parameter estimation are two effective methods to solve the problem of poor data reliability and data missing,but current research focuses on only one of the methods with the joint method less studied.Therefore,this study proposed an integrated framework of data reconcilation-parameter estimation-process simulation for physical inputs with significant errors or missing data in distillation systems.An accelerated algorithm was developed to accelerate and simplify the solution of the two-layer framework,and a steady-state simulation algorithm was designed considering convergence and efficiency,which was applied to the inner layer of the double-layer framework.In addition,wavelet transform was used to realize automatic stability judgment of original industrial data in continuous periods.Manually decision of steady state is no longer necessary,and the integrated framework can be automatically used to complete the calculation and analysis based on automatic dividing of steady state conditions.Finally,the platform was applied to a phenylenediamine distillation system,and the success rate of the integrated framework was 98%under all steady-state conditions,which proves that the framework and the algorithm have good applicability and convergence.